Research On Primary Selection System Of Black Tungsten Ore Based On Machine Vision

The primary selection of tungsten ore has the characteristics of improving the ore enrichment efficiency and reducing the production cost of the enterprise.Metallurgical enterprises have widely adopted the primary selection process,but the artificial primary selection of black tungsten ore has increased labor costs and increased the labor intensity of workers.With the extensive research and application of machine vision,the combination of industrial production and machine vision has been rapidly promoted,making industrial production develop in the direction of automation and intelligence.Therefore,the study of ore automation primary selection is of great significance and practical application value.This subject aims to solve the problem of primary ore selection,and combines with machine vision to realize the automatic primary selection of black tungsten ore.This topic first analyzes the production status of black tungsten ore in Gannan area,and puts forward the problems in the current primary ore selection process.Then,based on the analysis of the primary ore selection process requirements,the machine vision was used to sort the black tungsten ore,and the machine vision-based black tungsten ore preliminary separation system was designed,and its feasibility was verified.Finally,the function modules of the black tungsten ore primary selection system based on machine vision are determined,and the function modules are analyzed and selected,and the machine-based black tungsten ore primary selection system is established.According to the requirements of ore sorting process,a circular calibration plate was used to calibrate the industrial camera,and the image distortion was corrected,a mathematical model of the relationship between the image and the ore position was established,and the positioning and particle size measurement of the ore was realized,and the positioning accuracy and measurement accuracy were guaranteed.Through calibration experiments,the calibration accuracy of industrial cameras is verified.In order to make the gas injection port can accurately sort the black tungsten ore,the correspondence relationship between the black tungsten ore and the gas injection port is designed,and the mathematical model of the spatial position of the gas injection port and the position of the black tungsten ore is established.Identifying black tungsten ore is the core content of black tungsten ore primary selection system based on machine vision.In view of the ore image has the characteristics of dust and ore debris,and the complex color characteristics of the black tungsten ore,the ore image is gray-scaled and noise-reduced,so that the image has higher clarity.In order to accurately detect the ore position and particle size,threshold segmentation and edge detection-based segmentation methods are analyzed and studied.By analyzing the effect of ore segmentation,a Canny edge detection algorithm is proposed to segment the ore image.In view of the edge discontinuity in the edge detection segmentation method,a morphological connection edge contour is proposed to realize the connection of small neighborhood breakpoints,and the ore contour search method is improved,so that the contour search algorithm can accurately find the black tungsten ore contour.Through statistics of the color characteristics and particle size information of the ore,a threshold is set to analyze the tungsten content of the ore,and ore debris that does not require sorting is screened.Finally,the upper computer human-computer interaction software and the lower computer controller software based on ore positioning and particle size measurement,black tungsten ore segmentation and identification are designed.Based on OpenCV and QT,the developed host computer software has the characteristics of modularity,integrates functions such as debugging,operation and ore data statistics of industrial cameras and other peripherals,and has scalability.The lower computer software based on STM32 communicates with the upper computer in real time through the serial port,and cooperates with the upper computer to complete the ore sorting.By experiments,the primary selection system of black tungsten ore based on machine vision in this subject can accurately identify the black tungsten ore,locate the ore position and sort the black tungsten ore,which meets the process requirements of sorting black tungsten ore and has practical application value.